© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 70445
    Prev
    Next
    INDEX
    Explanations

    truth and concepts of correctness

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron strongly activates on occurrences of the word “truth” (and closely related terms like “true,” “false,” and “illusion”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    蚍
    -1.73
    
    -1.73
    囖
    -1.69
     huit
    -1.59
    辚
    -1.56
    érêt
    -1.55
     Inſ
    -1.51
     testen
    -1.48
    beutel
    -1.45
    喁
    -1.45
    POSITIVE LOGITS
     which
    1.88
    2
    1.60
    最初の
    1.58
     utilizar
    1.56
     Behaup
    1.55
    9
    1.52
     który
    1.49
     we
    1.48
     naturais
    1.47
    又は
    1.44
    Activations Density 0.014%

    No Known Activations